Add adaptive packet loss handling

docs/api-dotnet.md

@@ -228,8 +228,17 @@ moves members to Lobby (even if full), clearing their streams. Edits stop existi
 so clients must negotiate the updated audio configuration. Channel names/topics/passwords
 are limited to 128/4096/1024 UTF-8 bytes. Audio requires Opus, 48 kHz, mono/stereo,
 500–512000 bps, integral 5/10/20/40/60 ms frames and valid application/loss/complexity.
-Database v3 persists DRED with the rest of the channel audio configuration. Opening a v1 or v2
-database adds the DRED column with a disabled default before accepting channel updates.
+Database v4 persists DRED and `AudioConfig.packet_loss_mode` with the rest of the channel audio
+configuration. Opening older databases adds missing columns with disabled/manual defaults. The
+mode's zero value is manual for wire and database compatibility. Automatic Fast/Balanced/Stable
+modes use 3/10/30-second sender-uplink
+windows respectively. The server derives loss only from authenticated media counters, requires at
+least 20 expected packets, caps the applied Opus hint at 30%, and sends changed values to capable
+clients with `PacketLossUpdate`. The managed audio worker applies the mutable encoder control;
+device callbacks remain free of networking, allocation, and synchronization. The stored manual
+percentage is retained as the bootstrap value while an automatic window fills.
+Support is advertised as `adaptive-packet-loss` in both hello feature lists. Older senders keep
+using the bootstrap value, and edits from older administrators preserve an existing automatic mode.
 
 Session permissions gate kick/ban/move/mute and account operations. Only administrators
 can grant permissions; account-administration permission cannot grant administrator status.